Virtual Strain Gage (VSG) Size Calculation in Correlate 2019+

[[Da...]]

Hello,

Does anyone calculate a virtual strain gage (VSG) using parameters from Correlate 2019 and later?  

I have done this using parameters from Aramis V6.3, using Equation 7.2 in the iDIC best practices guide.  

L(window) is computation size, L(step) is step size, L(subset) is facet size. 

From what I understand, strain computation in 2019+ is computed using a weighted average/normal distribution along the strain window (tensor neighborhood).  Because of the weighted average, to me, VSG is not a 1:1 calculation anymore.  

There is the "reference length" feature in Correlate, but for a given change in tensor neighborhood, reference length does not change at the same rate.  

Is there an accepted method for calculated a VSG in GOM Correlate 2019+?

 

Thanks,

Dave

[[Ma...]]

Hi 

I think you may create a VSG easier in GC 2019+

1. Create an inspection you need - eg. EpsY

 

2. Click with LMB to inspection to activate it (EpsY is highlighted not Surf. Comp):

3. Press Ctrl RMB (activate I-Inspect and chose:

4. Chose area of interest 

 

(you may also use tools on the bottom:  for increasing or decreasing area size)

 

And after OK 

you will got a curve of AVG values from area (including time step etc due to Inspection of EpsY itself) 

 

Real strain gauge also do the same - averaging of strains below the strain gauge measuring area.

 

MK

[[Da...]]

Hi Marcin,

Thank you for your reply.  Your method of creating a VSG makes sense.  Creating an area that matches the "grid" of a typical strain gage is a smart way to do it.  

I do have one question.  Your method is averaging the strain over a defined area of your current surface component.  When you create that component, you set a strain tensor neighborhood size for the computation.  It is my understanding that the strain tensor neighborhood is the number of point distance steps (times 2) that are included in each strain tensor computation.  This means that adjusting the strain tensor neighborhood number effectively adjusts the "averaging" area that is used for the strain tensor calculation.  

So, if I was to use your method of creating an area that averages the strain over that given area, wouldn't I be taking an "average of an average", because tensor neighborhood is already an "averaging" method?

Thank you,

David Gottschalk  

[[Ma...]]

Here is a part from dic-strain-computation-basics.pdf (you should have access for this document afer logging in)

 

MK

[[Da...]]

Hi Marcin,

Yes, I have seen that document.  That is what I was referencing in my previous post.  As strain tensor neighborhood is increased, the software uses a larger and larger area to compute the strain.  In my mind, this is the "virtual strain gage" getting larger and larger. Therefore, using your method, you would be taking the average of a variable measurement, that depends on tensor neighborhood size.  

My understanding is that the total area under the strain tensor neighborhood does not directly reflect the size of a strain gage grid.  I think the algorithm GOM correlate uses is much more complex.  That is why I am trying to figure out a method to correlate the strain tensor neighborhood size (area in the images you have shared), to a theoretical strain gage grid size.  

[[Le...]]

Dave, 

Good question. One of the main takeaways I got from the iDICs VSG calculation is that it essentially boils down to "the pixel area from which is used to compute strain" which includes facet/subset size. This is my my gripe with GOM's "reference gage length" calculation: they provide a reference gage length that is independent of Facet/Subset size. 

I believe the iDICs VSG Equivalent equation applied to GOMs Correlate software would be the one I attached that's in green text, "VSG,A." 

In either case, neither definition is a 1:1 to a conventional foil strain gage. I never ran that sort of study, and perhaps someone else on the boards may have a better 1:1 VSG equation. 

Hope that helps, 
Levon